3D Object Proposals for Accurate Object Class Detection
3D Object Proposals for Accurate Object Class Detection
复制标题
DOI：
--
复制
发表时间：
2015-12
期刊：
影响因子：
--
通讯作者：
Xiaozhi Chen;Kaustav Kundu;Yukun Zhu;Andrew G. Berneshawi;Huimin Ma;S. Fidler;R. Urtasun
中科院分区：
文献类型：
--
作者：
Xiaozhi Chen;Kaustav Kundu;Yukun Zhu;Andrew G. Berneshawi;Huimin Ma;S. Fidler;R. Urtasun
本文的目标是在自动驾驶的背景下生成高质量的3D对象提案。我们的方法利用立体图像以3D边界框的形式放置建议。我们将问题描述为最小化一个能量函数，该能量函数编码对象大小先验、地平面以及几个深度信息特征，这些特征导致自由空间、点云密度和到地面的距离。我们的实验表明，在具有挑战性的Kitti基准测试中，与现有的RGB和RGB-D对象建议方法相比，性能有了显著的提高。结合卷积神经网络(CNN)评分，我们的方法在所有三个Kitti对象类上的性能都优于现有的所有结果。
The goal of this paper is to generate high-quality 3D object proposals in the context of autonomous driving. Our method exploits stereo imagery to place proposals in the form of 3D bounding boxes. We formulate the problem as minimizing an energy function encoding object size priors, ground plane as well as several depth informed features that reason about free space, point cloud densities and distance to the ground. Our experiments show significant performance gains over existing RGB and RGB-D object proposal methods on the challenging KITTI benchmark. Combined with convolutional neural net (CNN) scoring, our approach outperforms all existing results on all three KITTI object classes.